In order to realize auto guidance for attack UAV

a TV guiding system was designed. To solve the nonlinear filter problem in guidance system

the particle filter algorithm is proposed. A resampling algorithm was put forward after a careful analysis on the degeneracy phenomenon. The weight value of each sample was directly used to determine whether this sample should be kept or be discarded and how many times a sample should be reused. An adaptive threshold based on statistical analysis was employed here to classify the small weights and the large ones. A correlation tracking method based on particle filters was proposed. The experimental results indicate the guidance precision of new resampling algorithm excelled the old one .